Contesting Captions: Netflix,Fan Campaigns,and the Labor of Access

Abstract

This article theorizes the work of deaf fans and audiences in securing streaming video captions, which it names the labor of access. While prior scholarship on captions tends to emphasize policy and regulation, streaming video lacks the legislative framework for disability access seen with televisual captions in the United States. Drawing on disability studies and fan studies, this article shifts focus to fan campaigns, which become a central site for renegotiating industry accessibility practices in the absence of regulatory oversight. Using Netflix as a case study, it chronicles how audience campaigns contribute to both the company’s initial provision of streaming captions and the ongoing work of captioning quality control. It then argues that fan-driven access victories in the era of global streaming are uneven and contingent: a successful captioning campaign in one context has no necessary impact in others.

In the top right corner of Netflix’s video streaming interface, there is a flag icon that opens a menu for user-driven error reporting. If you encounter a problem while watching something, you can click the flag and choose your error out of a list of common failures Netflix provides. One provided error category is “subtitles and captions,” and choosing it reveals check boxes that name a variety of ways in which Netflix captions can fail: mistakes in spelling and transcription, mismatches with audio, problems with the display of text. Netflix’s error reporting mechanism suggests that captions are prone to failure—they break often enough to justify a streamlined pathway for identification and remedy—and, just as significant, that that the work of caption correction begins outside the company payroll. Via this user-facing menu, Netflix relies on its viewers to notice, describe, and report where and how captions go wrong.

Netflix’s reporting system thus reveals something important about captions in the era of streaming platforms: they depend on the audience. This article theorizes the work of d/Deaf ¹ fans and audiences in securing streaming video captions, which it names the labor of access. While prior scholarship on captions tends to emphasize policy and regulation, streaming video lacks the legislative framework for disability access seen with televisual captions in the United States. Drawing on disability studies and fan studies, this article shifts focus to fan campaigns, ² which become a central site for renegotiating industry accessibility practices in the absence of clear regulatory requirements. Using Netflix as a case study, it chronicles how fan campaigns contribute to both the initial provision of streaming captions and the ongoing work of captioning quality control. Though Netflix presents captions as a platform affordance—a designed feature it offers users—it is less a voluntary provider of access than a battleground for hard-won and often partial victories by audiences invested in assistive technology. Thus, we need to look to fan labor as the initial motive force behind accessibility gains.

In making this argument, I draw on fan accounts from Twitter, policy documents from Netflix and the U.S. government, and press coverage of streaming captions. Through combining these sources, I adopt what Ellcessor et al. (2017, 14) call an “integrated approach,” which entails “studying media texts not in relative isolation, but together with their industrial conditions of production, the social, political, and material contexts of their reception, and the active participation of audiences in producing meanings.” While an integrated approach resonates with many in cultural studies that highlight how media texts live in and move through the world, it has a particular investment in the actions and experiences of disabled communities. This investment undergirds my insistence on d/Deaf fans as the focal point of analysis. Instead of placing Netflix at the center of the story of digital captioning, I ask: how do fans invested in captions achieve access on Netflix? In excavating the history of contestation folded into platform function, I emphasize the agency of d/Deaf audiences and activist organizations, who use the language and strategies of fandom to win gains in an environment inhospitable to regulatory challenge.

This article consists of four sections. The first examines the regulatory model of disability media access, drawing out why a theoretical framework suited to broadcast television fails in relation to streaming video. I then propose “labor of access” as a term for the forms of fannish organization and publicity generation that replace formal legal challenges around accessibility in deregulated digital contexts. Sections two and three analyze key contestations in the implementation of captioning on Netflix. Section two looks to the 2012 fan campaign and federal court case (National Association of the Deaf et al. v. Netflix) that provoked Netflix to commit to captioning 100 percent of its video content. I show how, even in a moment of apparent legal victory, the limitations of regulatory power and the necessity of fan labor were already apparent. I then turn to a 2018 controversy about the captioning for Netflix original Queer Eye. In this case, an access campaign centered on the quality of captions rather than their simple presence; this example demonstrates how the labor of access continues even after the initial provision of captions. A final section draws out the limitations of fan campaigns, which achieve meaningful impact in relation to individual media texts and companies but struggle to scale within the wider landscape of global streaming.

Captions Beyond Regulation

Captions are textual representations of dialogue and other sounds in audiovisual media. They often resemble subtitles—text translating dialogue from one language to another—but they take d/Deaf viewers as their primary audience. Consequently, they offer a more holistic representation of sonic experience: a subtitle would not need to render the sound of a knock on a door but a caption might, especially if the knock is plot-relevant.

Scholarly engagement with captions centers on questions of access. While much disability research approaches media through representation, analyzing the images of disabled people that circulate in popular culture (Barnes 1992; Chivers 2011; Jeffress 2022; Norden 1994, Osteen 2008), a growing body of work looks instead to media technologies and infrastructures, interrogating whether and how people with disabilities find routes to participation in culture. Media scholar Ellcessor (2012, 2016, 2018a, 2018b, 2019) has been particularly influential in this turn toward access, and I follow her in defining it as “the ability of a person with one or more disabilities to make meaningful use of a media technology, whether through assistive technologies or through modification of mainstream technologies” (2016, 12). Captions are one of a number of assistive technologies that seek to increase access, alongside other digital accommodations like audio description (Kleege 2016) and physical accommodations like wheelchair ramps, curb cuts, and disabled parking permits. Access is, importantly, not a simple question of yes or no: the level of access an artifact, place, or technology affords varies between groups and situations, and is measured along a number of axes. For example, a website that is carefully designed to be accessible to color-blind users (incorporating colors that provide sufficient contrast to ensure readability in grayscale) might fail to include alternate (alt) text for its images. Consequently, it would be inaccessible to users of screen readers, who depend on alt text to translate visual media into machine readable form.

As an object of analysis, captions are complex because their implementation varies greatly across media industries. Captions can be always on (open) or turned on and off by a user (closed), hard coded into video or stored as a separate file, movable or placed in a static position, widely available or offered as token exceptions. It is thus important to treat captions as a heterogenous field rather than a single consistent object, and to specify medium, technology, and distribution context in their analysis. Film theaters, for example, might have select screenings with open captions, while television, DVDs and streaming video tend to offer closed captions. And there are technical distinctions at play even between captions that appear functionally identical to an end user. DVDs encode captions as video object (VOB) files, while much streaming video uses timed text markup language (TTML). A media text that moves from DVD to streaming will thus need a new—or at least newly encoded—set of captions, a fact that creates complications for both scholars and practitioners grappling with captions.

The most developed analyses of captioning at scale come in relation to television. Ellis (2019, 145) traces the evolving policy landscape around TV captioning in the United States, United Kingdom, and Australia, arguing that, “With each new technological change—be it the emergence of the internet or the switch from analogue to digital television—activists and volunteers have had to re-fight for access to captions and, in particular, for legislative change to facilitate this.” For Ellis (2019, 160), access battles iterate: captioning is a problem that needs to be solved again and again, with each new site of media consumption. However, she sees the same fundamental process operating across technologies. In the face of insufficient access, d/Deaf audiences raise complaints about a media industry’s exclusion of them, which ultimately leads to laws that force greater access. In this account, consumer activism takes legislation as its ultimate goal, for content providers act through “legislative determinism”: because they see disabled communities as outside their mainstream consumer base, they will only caption when a law requires them to, exactly as much as a law requires them to.

Ellis’ emphasis on consumer activism resonates with my project, but there are limitations to applying her model to streaming captions, which has a much less developed regulatory framework than television. Consider the divergence between laws governing streaming and ‘traditional’ television in the United States. A set of laws from the 1990s establish clear access expectations for TV. The Americans with Disabilities Act (1990) establishes a legal framework for access claims, specifically requiring that televised public service announcements caption any “verbal content.” The Television Decoder Circuitry Act (1990) requires that TV sets sold in the United States must come with the technical capability to decode and display the closed captions that accompany broadcast content. And the Telecommunications Act of 1996 gives the Federal Communications Commission the power to regulate captioning—to establish standards of quality and a date by which broadcast content must be fully accessible through closed captions. At this point, captioning becomes a generalized, legally mandated component of television broadcast, a feature that accompanies the vast majority of programs and which can be viewed on the vast majority of apparatuses through which people access them.

Streaming video, in contrast, has no captioning regulations that apply medium-wide. Rather, the 21st Century Communication and Video Accessibility Act (2010) routes online caption requirements through televisual appearance: it requires “the provision of closed captioning on video programming delivered using Internet protocol that was published or exhibited on television with captions.” That is, content that was captioned for television must also be captioned for digital distribution; networks like CBS and ABC must provide high-quality captions for content made available through their websites or streaming services like Hulu. However, video that is only available online—from Netflix originals to user-uploaded TikTok videos—is largely outside the purview of regulatory requirements regardless of its popularity. Legislative determinism suggests that this regulatory environment should lead to few if any streaming captions, and yet Netflix, Prime Video, Disney Plus and other prominent streamers provide captions for original and online-exclusive content, programming that has never and will likely never be subject to traditional TV broadcast.

How do we understand the disconnect between regulatory requirement and industry practice? One possibility is to expand a policy framework beyond legislation. Writing of the emergence of online captioning, Ellcessor (2011, 331) argues that the landscape of media access takes shape in relation to multiple interlocking policy systems: national, international, industrial, and voluntary. In addition to formal laws in the United States and abroad, she positions the internal decisions of corporations like Google as a form of industrial policy, which shapes “norms of corporate behavior and consumer expectations outside of explicit political processes” (Ellcessor 2011, 332). Applying this approach to Netflix would encourage us to look at its guidelines around captions, and to consider the market forces that make captions an attractive business choice even in the absence of firm legal requirements. However, such a move risks leading us away from the interests and practices of d/Deaf people. As Ellcessor (2011, 331) points out, corporate commitments to access are not necessarily commitments to disabled communities: many online captions are “motivated by the neoliberal business imperatives of the contemporary digital media industry” and “threaten to occur in the name of deaf and hard-of-hearing Americans while ignoring their specific needs”. To take Netflix policy at face value is thus to separate captions from the embodied community that they are ostensibly for.

As a remedy, I focus the accessibility work undertaken by d/Deaf fans. My approach draw on fan studies scholars who position media consumption as a form of labor (Busse 2015; De Kosnik 2012; Stanfill 2019; Stanfill and Condis 2014), one increasingly valuable and visible with the rise of digital platforms (Green & Jenkins 2009). Conceptualizing fandom as labor allows us to attend to how audiences work both with and against media industries: fans publicize and add value to corporate media properties, but they also agitate in opposition to industry practices felt to be unfair or exploitative. Fans mobilized as a collective are particularly successful at making an intervention into industry practice, and fan campaigns have undertaken goals from saving an imperiled show (Savage 2014), to contesting representational tropes for queer characters (Navar-Gill and Stanfill 2018), to critiquing media creators’ use of social media (Ravell 2023).

By recasting fan labor as the labor of access, I draw attention to ways d/Deaf fans and audiences mobilize the work of media consumption in the name of increased participation in media culture. ³ I imagine the labor of access as an expansive term that could be of use in analyzing and crediting any non-industry access work done in relation to media consumption, translation, and circulation. In explicating it here, I focus on fan captioning campaigns because they are a form of work originating outside the media industry but uniquely subject to cooptation by it. Netflix controls its own captioning practices, so fans cannot get captions on Netflix without its participation. ⁴ For this reason, once Netflix implements captioning improvements, it can and does take exclusive credit for changes that originate elsewhere. I thus use the labor of access as an analytic tool to get at the forms of fan work that are least visible. However, it could also be deployed in relation to a range of extra-industrial practices, including community captioning: the CaptionPlease subreddit, for example, allows d/Deaf people to request unofficial captions for specific YouTube videos they need or want access to, and YouTube once allowed community members to supply captions on other people’s videos (though the platform has since disabled that feature).

In the following sections, I place fan campaigns as a central mechanism for access interventions into industry practice in the absence of legislation. In the streaming era, we see direct negotiation between invested audiences and streaming companies around issues like captions, and fan demands can become encoded in corporate policy and practice in the wake of successful campaigns. Fan work thus becomes the labor of access: an ongoing, iterative practice that seeks to increase participation in the mediated sphere for d/Deaf audiences. Importantly, the labor of access happens via targeted campaigns, wherein fans make a specific ask of a specific media industry entity and then work to make that ask difficult to refuse. The fan captioning campaigns detailed in the next two sections are not unknown; they already appear in media access scholarship, as brief asides in accounts focused on policy (Ellis and Kent 2011, 139; Ellis and Kent 2015, 11; Ellcessor 2011, 330; Ellis 2020, 164). In revisiting them, my project is less about discovery than reorientation: I aim to decouple disability activism from a necessary end at legislation.

The Wizard of Oz and Caption Presence

When Netflix launched as a DVD-by-mail rental service, captions were already a widespread feature of this distribution technology: Netflix could simply buy captioned DVDs and rent them out, offering access without having to do additional work in service of it. In transitioning to streaming video in 2007, however, the company encountered a very different accessibility environment: online video was so new that there was no easy, pre-built captioning function that could be implemented at scale. In a 2009 blog post, Netflix Chief Product Officer Neil Hunt articulates the problem as an infrastructural one: Microsoft Silverlight, the software framework Netflix used to build its streaming video player, had “weak or nonexistent” support for captioning files, so Netflix could not offer closed captions. And always-on open captions would alienate hearing audiences: “The majority of viewers would object to English captions on English content, so we have to figure out how to let individual viewers turn them on and off” (Hunt 2009). Captioning is presented here as a desired but impractical and inessential feature, one that Netflix will implement in a vague future tense. Though comments on the blog post immediately contested its logic, pointing out that Silverlight did support caption files (Tim Heuer, June 12, 2009 8:29 PM) and that other streaming services already offered closed captions (Meg, June 12, 2009 9:17 PM), these complaints had relatively low visibility and impact. Netflix maintained its status quo—no captions—for the next few months.

Then Netflix announced a free, online screening of The Wizard of Oz set for October 3, 2009, in conjunction with the film’s 70th anniversary. Though Netflix subscribers received a premium, high-definition viewing experience, anyone with an internet connection could view The Wizard of Oz regardless of their subscription status in a move meant to drum up publicity for Netflix’s growing streaming service (Dentler 2009). At the time, DVDs were still Netflix’s primary source of income and subscribers, but streaming already occupied an important place in Netflix’s imagination of the future, and the company worked to make The Wizard of Oz a media event that would attract significant attention, even pairing it with a Central Park screening and concert (Hetrick 2009). In this way, Netflix’s presentation of The Wizard of Oz was a media event: it marked an escalation in visibility for the streamer, and an attempt to translate public affection for a classic film into an expanded subscriber base.

However, like all Netflix streaming content at the time, The Wizard of Oz would not have captions. If the lack of captions on Netflix’s wider library provoked a simmering discontent, The Wizard of Oz functioned as a flashpoint, sharpening access complaints into actionable anger. Deaf actress and activist Marlee Matlin describes the visceral charge of planning to watch with her children only to discover the absence of captions: “Right there, my daughter saw a barrier in front of my face. . . There was nothing I could do in this special moment. I was robbed of the access that I had for so long.” (ABC News 2014) As someone who played Dorothy onstage as a child, Matlin wanted to share this film with her family; the denial of access felt especially urgent because of her affective investment in it, and because Netflix marketed the movie as for everyone. Thus, while captions fail to exist for all streaming media on Netflix in 2009, attention and care is not evenly distributed among them. Netflix’s promotion of The Wizard of Oz set the stage for a targeted access campaign—one that would prove more successful than prior complaints about captioning.

In the weeks leading up to Netflix’s The Wizard of Oz showing, Matlin took to Twitter to express her dissatisfaction. She tweeted, “@netflix answer why Wizard of Oz wont [sic] be closed captioned I guess free for ALL except 30 million deaf people” and “@netflix is streaming “Wizard of Oz” free but no plans to close [sic] caption for 30 million deaf ppl Email [e-mail] Thanks!” (8 Sept 2009, @MarleeMatlin, Twitter). She calls on Netflix itself to fix the problem, and also asks for help in articulating grievance from others invested in captioning on The Wizard of Oz. This is a key component of the fan campaign: the power of an audience-based intervention into the operation of media industries comes from iteration, from the repetition of a desired outcome in so many voices that media executives have to grapple with it. And this plea does achieve repetition: actor Ashton Kutcher spreads Matlin’s request through his own Twitter following, many reach out to Matlin or Netflix directly on captions for The Wizard of Oz, and a letter-writing campaign culminates in the National Association of the Deaf (NAD) posting a public letter to Netflix. ⁵

Throughout the campaign, fannish investment in The Wizard of Oz is a repeated invocation. As the NAD write in their letter: “This classic movie appeals to a wide range of people—from children to senior citizens—many of whom have a hearing loss. . .This event provides a unique opportunity for Netflix to demonstrate its commitment to providing accessible entertainment to 36 million Americans who are deaf or hard of hearing. We want to be included in this 70th anniversary celebration of ‘The Wizard of Oz’ with our children and families.” (NAD 2009). Instead of asking for captions on all Netflix content, this is a campaign about a single media experience that has heightened significance to its potential audience. I make this point because the distinction between captioning one movie and captioning all content is useful in terms of strategy: Matlin, the NAD, and d/Deaf fans get to emphasize how little work it would be to create streaming captions for this one movie, which already exist in TV and DVD versions. In the past, Netflix emphasized the untenable labor requirements involved in creating captioned versions for its entire catalogue (Hunt 2009). This line of argument does not work in response to The Wizard of Oz: creating captions for a single film would not require an exorbitant investment of money or time.

Though The Wizard of Oz did not receive captions before its showing, Netflix did make captioning changes: it began captioning streaming content in April 2010 with an initial library of 100 captioned titles (Hunt 2010), and offered captions on 30% of its online video library by February 2011 (Hunt 2011). However, the lack of a firm promise to caption all content was disappointing, so activists turned to legal remedies. Along with the Western Massachusetts Association of the Deaf and Hearing Impaired, the NAD sued Netflix under the Americans with Disabilities Act (1990), arguing that Netflix is a place of public accommodation and so should be required to provide access via captions. The complaint makes specific mention of Netflix’s failure to caption The Wizard of Oz, reinforcing the importance of fan campaigns within broader disability activism. The case ended in a settlement in 2012, in which Netflix agreed to caption a hundred percent of its content library by 2014 (Mullin 2012).

While the NAD cast this case as a landmark precedent for digital access, its impact on the landscape of streaming captions is notably limited (NAD 2012). There are multiple factors restricting its reach. The first is the legal mechanism through which the settlement operated: it is a consent decree, which means that Netflix and the NAD came to a voluntary agreement instead of a legal body creating a rule with broad applicability. In other words, Netflix agreed to change its behavior as a company—to create captions where it previously had not—but nothing about the decision necessarily translates to any other streaming company. This agreement was also time limited: after an initial supervisory period of 3 years, the consent decree expired in 2016.

This marks a significant shift from legal decisions that hold sway over an entire medium to ones that hold over a particular platform or corporate entity for only a specified period of time. In previous eras, landmark decisions often applied to an entire category of media content; consider how civil rights activism achieved legal victories with systemic reach in relation to television (Classen 2004; Perlman 2016). In the digital era, convergence and multiplicity complicate media definitions. In the United States, there is so far no coherent legal category like “streaming platforms,” “digital video” or “web media.” It is not clear, after all, whether Netflix and YouTube are the same kind of thing; both traffic in online video, but YouTube allows for user submission where Netflix does not.

This case consequently has ambivalent results. Netflix as a technical system now enables captions, and as of 2024, Netflix captions almost all audiovisual streaming content available in the United States. Most properties premiere with captions (whether Netflix originals or content licensed from elsewhere), but Netflix has no legal requirement to continue providing them. Consequently, this moment is less a legal or regulatory change than an audience-driven renegotiation of industry captioning practice. And, as the next section will show, negotiations continue even beyond the provision of captions.

Queer Eye and Caption Quality

In June 2018, Netflix released the second season of Queer Eye (2018), an unscripted reboot of the earlier Bravo series Queer Eye For the Straight Guy (2003–2007). Netflix’s Queer Eye made headlines for its frank, intersectional portrayals of queerness, but also, unexpectedly, for its captioning (Allen 2018; Cooper 2018; Kilkenny and Shanley 2018). When the second season premiered, d/Deaf viewers noticed that not all spoken words appeared in the provided closed captions—lines of text meant to provide visual access to the audio track. Explicit language was the most noticeable omission, but there were also more general elisions and alterations. As one viewer with partial hearing loss details:

The first episode of the second season, “God Bless Gay,” was light on profanity, but I caught minor edits: Bobby would say, “In the beginning, I was honestly really apprehensive about this renovation. . .” but the captions would show, “In the beginning, I was apprehensive about this renovation.” In that same episode, Antoni tries a bite of egg salad and joyfully pronounces, “That’s freaking delicious!” but it becomes a bland, “That’s delicious” in the captions.

Finally, 2 minutes into the second episode, “A Decent Proposal,” Tan’s exclamation that the straight guy of the episode “Shit or get off the pot!” when it comes to proposing to his girlfriend is edited to, “Crap or get off the pot” (Ratcliff 2018).

Each individual change might seem insignificant, but taken together they alter the texture of conversations among the Queer Eye cast, flattening nuance and intensity. Thus, captions intended to provide access to the experience of Queer Eye instead misrepresent it.

A fan campaign grew up around Queer Eye’s captioning, as viewers took to Twitter to document, discuss, and lobby for fixes to captioning inaccuracies. For example, user MortuaryReport tweeted: “I really wish @netflix captions for #QueerEye2 weren’t bleeping profanity AND changing the profanity used in the captions. It’s really not awesome” (26 June 2018, Twitter). He then went on to detail the stakes of these changes: “it fundamentally changes the experience of the television show for anyone who is d/Deaf or HOH, and it does so without their consent. That’s seriously ableist, @netflix.” Another d/Deaf user is more pointed in their criticism, equating the changes with censorship: “Okay, @netflix. I want to know why you don’t caption every single word. I can see what people are saying not matching up with the captions. Also, pardon me, but DO NOT FUCKING CENSOR PEOPLE. I want to know what is said word. for. word.” (24 June 2018, @shan_no_says, Twitter).

Note that captioning complaints center on the quality of the captions rather than their simple presence or absence. It is tempting to imagine that if caption exists, they are good enough, for captioning as a process can seem easy, automatic, or outside purposeful human construction. However, it is a highly rhetorical act of writing. As disability scholar Sean Zdenek points out, there are many choices involved in the captioning process:

Which sounds are essential to the plot? Which sounds do not need to be captioned? How should genre, audience, context, and purpose shape the captioning act? What are the differences between making meaning through reading and making meaning through listening? Is it even possible, given the inherent differences between, and different affordances of, writing and sound, to provide the same information in writing as in sound? (Zdenek 2015)

For Zdenek, the presence of captions alone does not translate to the ability to meaningfully engage with a media text. He argues for evaluating captions in relation to sonic narrative experience: to use his example, captions should not reveal the speaker of a line to be a cannibal before the audio track has made this fact clear to hearing audiences. When viewers object to Netflix’s erasure of profanity, they are making a related complaint: user MortuaryReport wrote that altering or removing profanity fundamentally changed the show. For this viewer, it is not enough to know that there was profanity, a fact that bleeped out text can communicate. What was said and how many times it was said provides valuable insight into the affective texture of the show and to the interactions depicted within it. The profanity is, in other words, essential to the experience. Removing it enacts an infantilization of d/Deaf viewers, as Netflix decides on their behalf what kind of language is and isn’t appropriate for consumption via caption.

These tweets and others like them accrue thousands of likes and retweets, and eventually draw the attention of Karamo Brown, one of the Queer Eye “Fab 5”—the show’s five central personalities. Brown himself tweeted: “Reading everyone’s comments breaks my heart. I don’t know how much power I have but know, the next time I’m at Netflix I’m going to bring up this issue internally and won’t stop until something changes. Deaf and HOH people should have the same experience as everyone else! #TypoFixed” (28 June 2018, @Karamo, Twitter). Brown explicitly quotes fan comments, crediting them with drawing his attention to amplify the fan campaign. The next day, Netflix addressed Queer Eye captioning concerns from its customer service Twitter handle, admitting fault and promising a remedy: “We’ve heard about the caption issues on the service, specifically for @QueerEye. After looking into it, there’s a lot of dialogue missing from the Fab 5 that shouldn’t be. We’re fixing it. In some cases, we do bleep incidental profanity from our unscripted series.” (28 June 2018, @NetflixHelps, Twitter.) Netflix then thanked Brown and fans for helping bring their attention to the issue, emphasizing the vital role of fan labor in remedying captioning inadequacies.

Note that the entity fans and stars lobby is Netflix itself: the courts have dropped from the equation entirely, for Netflix gets to set its own guidelines and best practices for captioning. Netflix’s confidence in own autonomy in clear in its response to complaints. Netflix acknowledges that it made a mistake in captioning this particular show, but then doubles down on its own ability to set the rules: “In some cases, we do bleep incidental profanity from our unscripted series.” The divisions between incidental and purposeful profanity are not clear, so Netflix essentially leaves itself space to continue doing whatever it wants. Netflix promises to change because viewers got sufficiently loud—and drew the attention of someone sufficiently important in a Queer Eye star—but no law forces it to. And there is no guarantee that any other show will benefit from the extra attention given to Queer Eye captions: though this fan campaign was successful, its success was narrow, changing Netflix’s captioning practice only in relation to a single show. Queer Eye is a high-profile Netflix original, with a vocal fan base and a star who is attentive to online discussion. A different series might not receive the same level of responsiveness to captioning complaints. This dynamic leads to a world in which only the most popular shows, or those with the most committed viewers, or those whose stars happen to pay attention to fans get receive high quality captions. ⁶

Networked fan campaigns demonstrate both the power and precarity of digital labor of access; they can succeed in intervening into industry accessibility practices like captioning, but without the enforcement mechanism built into legal mechanisms they lack sector-wide impact. Thus, the labor of access makes for an uneven patchwork wherein not only do some platforms caption better than others, but even within the same platform some media properties receive better captions than others.

Conclusion

On a website providing guidelines for content creators, Netflix places great importance on creating caption files. However, after a token mention of access for d/Deaf audiences, Netflix goes to great lengths to demonstrate the value of captions for the hearing:

With the rapid growth of mobile devices, living rooms are no longer needed to enjoy your favorite film or television show. NETFLIX affords its members the ability to watch content on their own schedule, and subtitles allow members to enjoy content without the need for silence or headphones. Timed text files are paramount not only to the NETFLIX service, but also for the consumption of media in a world without the living room as the primary center of entertainment. (Netflix 2023)

Since Netflix enables people to watch content anytime and anywhere, on phones and laptops as well as televisions, environmental conditions make visual access to sonic content useful for everyone: you the generalized user may find yourself in a space that is too loud to hear well, or without headphones in a situation where playing sound aloud would disturb others. Netflix here makes a universalizing argument for captions: everyone is a potential beneficiary of them because media consumption has left the living room. By foregrounding changes in when and how people watch, Netflix imagines limitations to access as temporary, situational, and environmental. The particular important of access for disabled audiences disappears here, as does the history of fan advocacy that brought captions to Netflix in the first place.

Netflix’s narration underscores the stakes of unearthing the labor of access: if scholars do not seek out and highlight the work of disabled people, platforms will write access into their own stories on their own terms. Though corporations are slow to invest money and time in access projects, they are quick to take credit for them. There is political and economic value in presenting as a champion of access, even—or perhaps especially—if a corporation has once faced public criticism for accessibility failures. Thus, we should interrogate all corporate investments in access, and adopt methodologies that recenter the actions and experiences of disabled people.

In this article, I focused on the labor of access as a way to center the work of disabled audiences in achieving captions. I suggest that, while landmark legislation was once the barometer for measuring access gains, fan campaigns now aim instead at changing the accessibility policies and practices of individual companies. This shift makes for an environment where audience collectives wield the threat of negative publicity—a motivator that can be powerful, but has clear limitations. Their victories are uneven and contingent, for a successful captioning campaign in one context has no necessary impact on others.

Future research might turn the lens of fan labor on an expanded range of sites: How do fan campaigns operate in relation to other assistive technologies like audio description and capti-view devices? What forms of in-person organization and campaigning complement the digital labor described here? How does fan labor operate in other national contexts with distinct industrial and policy apparatuses? In approaching any of the questions, we focus on individual companies and platforms as the site of contestation. While accessibility regulation may someday create an umbrella category for online entertainment, that day is unlikely to arrive soon. In the meantime, there is not one overarching battle for digital access but rather a proliferating array of sites of contestation, each with its own logic and goals.